Ecological environment quality assessment method and device combined with multi-source remote sensing data

Abstract

The application discloses an ecological environment quality evaluation method and device combined with multi-source remote sensing data, relates to the technical field of remote sensing monitoring and ecological environment evaluation, and comprises the following steps: acquiring multi-source remote sensing data containing ecological environment characteristics; pre-processing the multi-source remote sensing data; extracting multi-dimensional ecological environment characteristics from the pre-processed multi-source remote sensing data; constructing a regional ecological characteristic vector based on the multi-dimensional ecological environment characteristics; inputting the regional ecological characteristic vector into a pre-constructed and trained ecological environment quality evaluation model group to obtain ecological environment quality evaluation results of multiple ecological divisions; and spatially splicing the ecological environment quality evaluation results of the ecological divisions to generate target ecological environment quality evaluation results. The method can realize continuous, fine and comprehensive evaluation of ecological environment quality.

Description

Technical Field

[0001] This invention relates to the field of remote sensing monitoring and ecological environment assessment technology, and in particular to a method and apparatus for assessing ecological environment quality by combining multi-source remote sensing data. Background Technology

[0002] Ecological and environmental quality is an important comprehensive indicator for measuring the level of regional sustainable development, playing a fundamental supporting role in macro-level decision-making such as national ecological civilization construction, resource and environmental supervision, and territorial spatial planning. Traditional ecological and environmental quality assessment mainly relies on ground monitoring station observations and quadrat surveys. Although these methods have high accuracy at the local scale, they are limited by factors such as the limited distribution of monitoring points, high survey costs, and insufficient timeliness, making it difficult to meet the needs of large-scale, long-term dynamic monitoring and comprehensive assessment of ecological and environmental quality.

[0003] With the continuous operation and data accumulation of Earth observation satellites, remote sensing technology provides a data foundation for ecological environment quality assessment that is timely, widely covered, and repeatable. Ecological environment quality assessment methods based on remote sensing ecological indices have received widespread attention; however, existing studies mostly rely on single or homogeneous optical remote sensing data, making long-term continuous monitoring difficult in cloudy and rainy areas. Although some studies have begun to explore the introduction of radar and optical data fusion, deep learning-driven data completion methods, and multivariate feature construction strategies, the potential of multi-source remote sensing data for the collaborative utilization of ecological information has not yet been fully explored.

[0004] Furthermore, existing ecological environment quality assessment methods typically employ fixed or globally uniform weights during index construction, neglecting the differences in dominant ecological processes and environmental response mechanisms among different ecological zones. This results in insufficient applicability and ecological interpretability of the assessment results at the regional scale. Purely data-driven weight inversion methods, under conditions of uneven sample distribution or noise interference, are prone to producing weight configurations that contradict ecological understanding, thereby affecting the stability and reliability of ecological environment quality assessment results.

[0005] Therefore, there is an urgent need for an ecological environment quality assessment method that can fully integrate the advantages of multi-source remote sensing data while taking into account regional differences. Summary of the Invention

[0006] The purpose of this invention is to provide a method and apparatus for assessing the quality of the ecological environment by combining multi-source remote sensing data, which can solve the above-mentioned problems existing in the prior art.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a method for assessing ecological environment quality by combining multi-source remote sensing data, wherein the method includes: Acquire multi-source remote sensing data containing ecological and environmental characteristics, wherein the multi-source remote sensing data includes: optical remote sensing data, radar remote sensing data, and topographic data; The multi-source remote sensing data is preprocessed; Multidimensional ecological and environmental features are extracted from preprocessed multi-source remote sensing data. These features include vegetation ecological features, water body and humidity ecological features, thermal environment ecological features, and soil and surface disturbance ecological features. Construct a regional ecological feature vector based on the aforementioned multidimensional ecological environment characteristics; The regional ecological feature vector is input into a pre-constructed and trained ecological environment quality assessment model set to obtain ecological environment quality assessment results for multiple ecological zones; wherein, the ecological environment quality assessment model set includes multiple ecological zone sub-models constructed with adaptive ecological environment quality weights for each ecological zone; The ecological environment quality assessment results of each ecological zone are spatially stitched together to generate the target ecological environment quality assessment result.

[0008] This invention also provides an ecological environment quality assessment device that combines multi-source remote sensing data, wherein the device includes: The acquisition module is used to acquire multi-source remote sensing data containing ecological and environmental characteristics, wherein the multi-source remote sensing data includes: optical remote sensing data, radar remote sensing data, and terrain data; The preprocessing module is used to preprocess the multi-source remote sensing data; The extraction module is used to extract multidimensional ecological and environmental features from preprocessed multi-source remote sensing data. The multidimensional ecological and environmental features include: vegetation ecological features, water body and humidity ecological features, thermal environment ecological features, and soil and surface disturbance ecological features. The construction module is used to construct a regional ecological feature vector based on the multidimensional ecological environment characteristics; The evaluation module is used to input the regional ecological feature vector into a pre-constructed and trained ecological environment quality evaluation model group to obtain ecological environment quality evaluation results for multiple ecological zones; wherein, the ecological environment quality evaluation model group includes multiple ecological zone sub-models constructed with adaptive ecological environment quality weights for each ecological zone; The generation module is used to spatially stitch together the ecological environment quality assessment results of each ecological zone to generate the target ecological environment quality assessment result.

[0009] This invention also provides an electronic device, comprising a processor, a communication interface, a memory, and a communication bus. The processor, communication interface, and memory communicate with each other via the communication bus. The memory stores computer programs. When the processor executes the program stored in the memory, it implements any of the above-described methods for assessing ecological environment quality using multi-source remote sensing data.

[0010] This invention discloses an ecological environment quality assessment scheme that combines multi-source remote sensing data. The scheme acquires multi-source remote sensing data containing ecological environment characteristics; preprocesses the multi-source remote sensing data; extracts multi-dimensional ecological environment characteristics from the preprocessed data, such as vegetation ecological characteristics, water body and humidity ecological characteristics, thermal environment ecological characteristics, and soil and surface disturbance ecological characteristics; constructs a regional ecological feature vector based on these multi-dimensional ecological environment characteristics; inputs the regional ecological feature vector into a pre-constructed and trained ecological environment quality assessment model group to obtain ecological environment quality assessment results for multiple ecological zones; and spatially stitches the ecological environment quality assessment results of each ecological zone to generate the target ecological environment quality assessment result. This invention, through techniques such as consistent preprocessing of multi-source remote sensing data, extraction of multi-dimensional ecological environment characteristics, and ecological quality assessment based on regional adaptive ecological environment quality weight modeling, not only integrates multi-source remote sensing data but also takes into account regional differences, enabling continuous, refined, and comprehensive assessment of ecological environment quality, thereby providing scientific support for ecological environment protection and resource management. Attached Figure Description

[0011] Figure 1 This is a flowchart illustrating the steps of an ecological environment quality assessment method combining multi-source remote sensing data according to an embodiment of this application. Figure 2 This is a structural block diagram illustrating an ecological environment quality assessment device that combines multi-source remote sensing data according to an embodiment of this application. Detailed Implementation

[0012] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0013] To overcome the shortcomings of existing ecological environment quality assessment schemes, such as relying on a single data source and lacking regional adaptability, this application provides an ecological environment quality assessment scheme that combines multi-source remote sensing data. This falls within the technical category of interdisciplinary applications of ecological environment monitoring, natural resource management, and geographic information science. Based on the comprehensive requirements of ecological environment quality assessment regarding ecological characteristics, surface conditions, and environmental background, this scheme uses optical remote sensing data, radar remote sensing data, and topographic data as multi-source remote sensing data for ecological environment quality assessment. First, consistency preprocessing of the multi-source remote sensing data is performed to obtain the data foundation for ecological environment quality assessment. Then, a multi-dimensional ecological feature system is inverted and constructed based on the processed remote sensing data, providing a feature basis for ecological environment quality index modeling in subsequent ecological environment quality assessments. Finally, the ecological environment quality index M-RSEI is constructed using the multi-dimensional ecological feature set partitioning for ecological environment quality calculation and mapping.

[0014] The following description, in conjunction with the accompanying drawings, details the ecological environment quality assessment method combining multi-source remote sensing data provided in this application through specific embodiments and application scenarios.

[0015] As attached Figure 1 As shown in the figure, the ecological environment quality assessment method combining multi-source remote sensing data in this application includes the following steps: Step 101: Obtain multi-source remote sensing data containing ecological and environmental characteristics.

[0016] The multi-source remote sensing data includes optical remote sensing data, radar remote sensing data, and topographic data.

[0017] The complementarity of multi-source remote sensing data in imaging mechanisms such as optics, radar, and thermal infrared enables it to comprehensively characterize vegetation growth, water features, surface structure, and surface energy exchange processes from different perspectives, providing key technical support for regional-scale ecological environment quality assessment. Based on remote sensing data, related research has gradually constructed an ecological monitoring system centered on vegetation indices such as NDVI (Normalized Difference Vegetation Index) and EVI (Enhanced Vegetation Index), and derived various comprehensive ecological environment quality assessment models. These methods typically achieve quantitative characterization of ecological environment quality by inverting key ecological parameters such as vegetation physiological parameters, surface temperature, and soil moisture.

[0018] Optical remote sensing data can include, but is not limited to, Sentinel-2 MSI (10m), Landsat-8 / 9 OLI (30m), Landsat TIRS (100m), and GF-6 WFV imagery, used to extract parameters such as vegetation index, humidity index, soil bareness, and spectral characteristics. Land surface temperature can be retrieved using Landsat TIRS (100m) to provide a basis for assessing drought stress and thermal conditions.

[0019] Radar remote sensing data may include, but is not limited to: Sentinel-1 radar imagery (SAR, 10m). Radar remote sensing data compensates for the optical limitations caused by clouds and rain, and is used to estimate surface structure information, humidity conditions, and backscattering characteristics of forest and grassland cover.

[0020] Topographic data can be used to extract parameters such as slope, aspect, and elevation from a DEM (ALOS AW3D, 30m), providing topographic constraints for topographic correction and quality assessment models.

[0021] In addition, this application may also obtain auxiliary data and sample data, including but not limited to: land use data, field ecological survey sample points, vegetation cover sample points, soil moisture and environmental observation station data, etc., for training and validation of the ecological environment quality assessment model group.

[0022] Step 102: Preprocess the multi-source remote sensing data.

[0023] To ensure the consistency of multi-source remote sensing data, the data is preprocessed in this embodiment to obtain the data basis for ecological environment quality assessment.

[0024] An optional method for preprocessing multi-source remote sensing data may include the following sub-steps: Sub-step 1: Perform radiometric calibration and atmospheric correction on optical remote sensing data containing ecological and environmental characteristics to convert optical remote sensing images into surface reflectance data.

[0025] Sub-step 2: Perform geometric correction and projection unification processing on the optical remote sensing data to ensure that the optical remote sensing images are consistent in spatial location.

[0026] Sub-step 3: Perform terrain correction on the optical remote sensing data based on the terrain data, and perform spatial resolution unification processing on the terrain-corrected optical remote sensing data.

[0027] In practical implementation, a feasible preprocessing method for optical remote sensing data may include, but is not limited to: radiometric calibration and atmospheric correction, geometric correction and projection unification, terrain correction, and spatial resolution unification. Specific preprocessing operations can be as follows: 1) Radiation calibration and atmospheric correction Sentinel-2 MSI: Atmospheric correction can be performed using the Sen2Cor tool; Landsat-8 / 9 OLI: Landsat Level-2 Collection 2 can be used; GF-6 WFV: Atmospheric correction can be performed using the FLAASH model. The FLAASH (Fast line-of-sight atmospheric analysis of spectral hypercubes) model is an atmospheric correction module suitable for correcting hyperspectral and multispectral remote sensing data in the visible to shortwave infrared bands.

[0028] After correction, all optical remote sensing data, also known as optical remote sensing images, are converted into surface reflectance data.

[0029] 2) Geometric correction and projection unification Multi-source optical remote sensing data are projected onto the same coordinate system, and sub-pixel-level geometric registration based on feature point matching is used to eliminate geometric deviations between different sensors, ensuring the consistency of images in spatial location.

[0030] 3) Terrain Correction The terrain correction step eliminates the interference of non-terrain features on spectral and scattering characteristics. Terrain correction was performed on Sentinel-2, Landsat-8 / 9, and GF-6 models based on DEM data and the C-correction method. The C-correction method, also known as the curve correction method, is a curve straightening method.

[0031] 5) Unified spatial resolution of optical remote sensing data In practical implementation, taking Sentinel-2 MSI (10m), Landsat-8 / 9 OLI (30m), Landsat TIRS (100m) and GF-6 WFV imagery as examples, in order to ensure the spatial consistency of subsequent feature fusion and ecological index calculation, the spatial resolution of the optical remote sensing data needs to be resampled to be consistent with that of Sentinel-2 MSI (10m).

[0032] The visible light spatial resolution of Landsat-8 / 9 OLI is 30m, which is adjusted to 10m using bilinear interpolation; the spatial resolution of TIRS is 100m, which is adjusted to 10m using bilinear interpolation; and the spatial resolution of GF-6 WFV is 16m, which is adjusted to 10m using bicubic interpolation.

[0033] Sub-step 4: Detect and repair optical pixels contaminated by clouds or shadows in radar remote sensing data containing ecological and environmental characteristics.

[0034] Preprocessing of the radar remote sensing data included: Sentinel-1 employs C-band dual polarization and is sensitive to vegetation structure, humidity, and surface roughness. Sentinel-1 Level-1 GRD data, i.e., radar remote sensing data, was downloaded from the Google Earth Engine (GEE) platform. This data had already undergone preprocessing by the GEE platform, including orbit file application, boundary noise removal, thermal noise removal, radiometric calibration, and terrain correction. The final image was expressed in decibels.

[0035] To ensure the integrity and stability of subsequent ecological feature extraction, optical pixels contaminated by clouds or shadows are detected and repaired.

[0036] Specifically, the FMask 4.0 algorithm can be used to identify cloud, thin cloud, cloud shadow, and snow pixels in remote sensing data acquired by Sentinel-2, Landsat-8 / 9, and GF-6. FMask, based on a combination of multispectral thresholding and thermal information, effectively distinguishes cloud bodies, shadows, and ground surfaces. This can be achieved using the following formula: For in Pixels that are constantly identified as clouds or shadows The optical observation value is considered to be missing. For this pixel, surface structure constraint information in the time and spatial dimensions is used for repair.

[0037] In the temporal dimension, cloud-free observations of the same pixel in multi-temporal optical imagery are used to construct temporal statistical features for preliminary estimation of missing pixels. This estimation employs robust statistics to reduce the impact of outliers on the results.

[0038] In the formula, This indicates a robust statistic, which is a weighted average. It represents the effective time phase set within the same historical period and its neighborhood.

[0039] Since radar imagery is insensitive to cloud and fog conditions and can stably reflect surface structure features, a set of pixels with similar radar backscattering characteristics to the target pixel is selected within the spatial neighborhood. By statistically analyzing the optical reflectance characteristics of these similar pixels at the same time, an estimation of optical information under spatial consistency constraints is obtained. Therefore, a pixel structure is constructed within the spatial dimension of the surface structure constraint. spatial neighborhood Filtering and pixels A set of pixels with similar radar backscattering characteristics: In the formula, For the same land category Inner pixel Standard deviation of backscattering in the neighborhood: Among them, land category information is obtained through land use data.

[0040] Repaired pixels reflectivity value for: In the formula, These are the weighting coefficients for the temporal and spatial constraints of the Earth's surface structure, with values ​​ranging from [0,1]. When multi-temporal data is abundant, Take the larger value; when continuous cloud cover is severe, increase the spatial repair weight of surface structure constraints.

[0041] It should be noted that the preprocessing process for optical remote sensing data, radar remote sensing data, and terrain data is not limited to the order specified in the above optional embodiments. The three types of data can be preprocessed in parallel or in any other combination of execution orders. In this application embodiment, no specific restrictions are placed on the preprocessing order of the three types of data.

[0042] Step 103: Extract multidimensional ecological and environmental features from the preprocessed multi-source remote sensing data.

[0043] Ecological environment quality is the product of the combined effects of multiple ecological factors, and a single remote sensing indicator is insufficient to comprehensively depict the regional ecological state. Taking into account the structural and functional characteristics of the ecosystem, and using preprocessed, temporally continuous, and spatially consistent multi-temporal remote sensing observation data, this application constructs an ecological environment characteristic system from four dimensions: vegetation status, water conditions, thermal environment characteristics, and soil and surface disturbance. Multi-dimensional ecological environment characteristics are extracted from the preprocessed multi-source remote sensing data, i.e., multi-source remote sensing data inversion is performed. The results of the multi-source remote sensing data inversion are used as the basic input for subsequent ecological environment quality assessment, thereby avoiding biases in ecological environment quality evaluation caused by a single observation dimension.

[0044] Among them, the multidimensional ecological environment characteristics include: vegetation ecological characteristics, water body and humidity ecological characteristics, thermal environment ecological characteristics, and soil and surface disturbance ecological characteristics.

[0045] In one optional embodiment, the method of extracting multidimensional ecological and environmental features from preprocessed multi-source remote sensing data may include the following operations: 1) Extract vegetation coverage and vegetation growth status from the preprocessed multi-source remote sensing data as vegetation ecological characteristics.

[0046] This operation involves remote sensing inversion of vegetation ecological characteristics. Vegetation is a core characterizing element of ecological environment quality; its coverage, growth activity, and structural characteristics directly reflect the health and stability of the ecosystem. In this optional embodiment, vegetation ecological characteristics are extracted from two aspects: vegetation cover (FVC) and vegetation growth status. Vegetation growth status may include, but is not limited to, NDVI and EVI.

[0047] 2) For different underlying surface conditions, water body identification information and surface humidity are extracted from the preprocessed multi-source remote sensing data as ecological characteristics of water bodies and humidity.

[0048] This operation involves the inversion of water body and humidity ecological characteristics. Water body and surface humidity play a crucial role in regulating the regional thermal environment and maintaining ecosystem stability. For different underlying surface conditions, water body and humidity ecological characteristics are constructed from two aspects: water body identification (MNDWI) and surface humidity characterization.

[0049] 3) Extract land surface temperature from preprocessed multi-source remote sensing data as a thermal environment ecological feature.

[0050] This operation involves the inversion of thermal environmental ecological characteristics. The surface thermal environment is an important indicator of the intensity of human activities and ecological degradation, significantly impacting the structure and function of ecosystems. The land surface temperature (LST) is retrieved using thermal infrared remote sensing data as a characteristic of the thermal environment's ecological features.

[0051] 4) Extract surface aridity and bare land index from the preprocessed multi-source remote sensing data as ecological characteristics of soil and surface disturbance.

[0052] This operation involves the inversion of ecological characteristics of soil and surface disturbance. Soil exposure level and the intensity of human disturbance are important indicators for measuring ecological degradation and land use pressure. This is achieved by constructing aridity... The bare land index (NDBSI) reflects the ecological characteristics of soil and surface disturbance.

[0053] The extracted multidimensional ecological and environmental characteristics used for ecological and environmental quality assessment are detailed in Table 1.

[0054] Table 1 Step 104: Construct regional ecological feature vectors based on multidimensional ecological environment characteristics.

[0055] The regional ecological feature vector constructed based on the above-mentioned inverted multidimensional ecological environment characteristics can be as follows: In one optional embodiment, when constructing a regional ecological feature vector based on multidimensional ecological environment features, the multidimensional ecological environment features can be standardized; the regional ecological feature vector is constructed based on the standardized multidimensional ecological environment features.

[0056] After standardizing the multidimensional ecological and environmental characteristics, a regional ecological feature vector is constructed. This vector serves as the input for subsequent principal component analysis and regional adaptive weight model, which can eliminate differences in the dimensions and numerical ranges of different features.

[0057] Step 105: Input the regional ecological feature vector into the pre-constructed and trained ecological environment quality assessment model group to obtain the ecological environment quality assessment results of multiple ecological zones.

[0058] Among them, the ecological environment quality assessment model group includes multiple ecological sub-models constructed by adaptive ecological environment quality weights for ecological zones.

[0059] Unlike existing ecological index models that use fixed or global weights, this application's embodiments employ regional adaptive weight modeling, enabling the ecological environment quality index to reflect the differences in dominant ecological processes across different ecological zones. To address the differences in dominant ecological environment factors under different land use types, a zonal weight modeling mechanism is introduced to dynamically adjust the contribution of ecological characteristics within different ecological functional zones, thereby improving the regional adaptability and ecological interpretation of the ecological environment quality assessment results.

[0060] In one optional embodiment, the ecological environment quality assessment model set is trained and generated in the following manner, specifically including the following sub-steps: Sub-step 1: Based on land use data and ecological function zoning results, the study area is divided into multiple ecological zones.

[0061] Sub-step 2: Construct a sample set in each ecological zone, with multidimensional ecological environment characteristics as input and historical ecological environment quality reference values ​​as output.

[0062] A feasible approach to ecological zoning and sample construction is as follows: Due to significant differences in natural background, intensity of human activities, and dominant ecological processes among different ecological zones, in order to reveal the heterogeneity of ecological response mechanisms within these regions, the study area was divided into zones based on land use data and ecological function zoning results. Divided into One ecological zone: In the formula, Indicates the first Each ecological zone.

[0063] In each ecological zone, a sample set is constructed, using multi-source remote sensing ecological features as input and historical ecological environment quality reference values ​​as output: In the formula, For cell position; , indicating in a pixel Normalized extraction Ecological characteristics (NDVI, EVI, etc.); This represents the reference value for ecological environment quality calculated based on the classic ecological index RSEI.

[0064] Sub-step 3: In each ecological zone, extract the first preset number of principal component features from the multidimensional ecological environment features based on principal component analysis.

[0065] A feasible method for extracting ecological gradient features based on principal component analysis can be as follows: Since there are correlations among multidimensional ecological features, direct modeling can easily lead to unstable weights and multicollinearity. Therefore, in this embodiment, principal component analysis (PCA) is performed on the ecological feature matrix in each ecological zone to extract the dominant ecological change gradient.

[0066] Ecological zoning The characteristic matrix in After standardization, its covariance matrix is ​​calculated and eigenvalue decomposition is performed: In the formula, Indicates ecological zoning The number of samples in the sample; Indicates ecological zoning The Middle Eigenvalues ​​of each principal component; It is the corresponding feature vector.

[0067] Principal component scores Represented as: Select the top contributors whose cumulative contribution rate meets the threshold condition. Principal components: in, Typically, 85%-90% is used.

[0068] Sub-step 4: Establish a regression model based on principal component features within each ecological zone.

[0069] A feasible way to build a regression model is as follows: To characterize the relative contributions of principal components to ecological environment quality within different ecological zones, this application establishes regression models based on principal component characteristics within each ecological zone: In the formula, Ecological zoning The intercept term; Principal components In ecological zoning The regression coefficients within. This is the sign function used to constrain regression coefficients. When ecological zoning... When the dominant factors within the ecosystem are positively correlated with the quality of the ecological environment (see Table 1 for details on the directional relationship between ecological factors and the quality of the ecological environment). If there is a negative correlation, . This represents the random error term. Furthermore, to avoid overfitting, ridge regression is preferred for constraining the coefficients.

[0070] Sub-step 5: For each regression model, map the regression coefficients of the regression model to normalized weights, and perform soft constraint correction on the normalized weights.

[0071] Regression coefficients reflect the relative strength of the principal components' influence on changes in ecological and environmental quality, but their values ​​are affected by the dimensions and sample distribution. To construct a weighting system that can be used for comprehensive evaluation of ecological and environmental quality, regression coefficients can be mapped to normalized weights: To avoid ecologically inappropriate results from purely data-driven weight learning under conditions of sample imbalance or noise, an expert constraint based on ecological consensus is introduced. This constraint modifies the regression results by limiting the relative importance range of dominant factors within different ecological zones and the reasonable range of weight variation (not allowing a single factor weight to approach 1), thereby enhancing the model's ecological interpretability and stability while maintaining its flexibility. The upper and lower bounds are confidence intervals derived from regression results of data from multiple periods in different ecological zones.

[0072] Sub-step 6: Within each ecological zone, construct a regional adaptive ecological environment quality index based on the normalized weights corrected by soft constraints.

[0073] In practical implementation, a regional adaptive ecological environment quality index can be constructed within each ecological zone: Step 106: Spatially stitch together the ecological environment quality assessment results of each ecological zone to generate the target ecological environment quality assessment results.

[0074] Spatially stitching together the environmental quality assessment results of different ecological zones yields an overall ecological environment quality distribution map of the study area. In practical implementation, the EEQI can be divided into several levels, such as excellent, good, medium, and poor, forming an ecological environment quality level map. It should be noted that the levels are not limited to the four levels of excellent, good, medium, and poor mentioned above; they can also be divided into 3 levels, 5 levels, etc. The specific rules for level division in this application embodiment are not specifically limited.

[0075] The target ecological environment quality assessment results can be ecological environment quality level maps, textual information used to characterize ecological environment quality, or a combination of textual and graphic information used to characterize the ecological environment.

[0076] The ecological environment quality assessment method combining multi-source remote sensing data disclosed in this application involves: acquiring multi-source remote sensing data containing ecological environment characteristics; preprocessing the multi-source remote sensing data; extracting multi-dimensional ecological environment characteristics from the preprocessed multi-source remote sensing data, such as vegetation ecological characteristics, water body and humidity ecological characteristics, thermal environment ecological characteristics, and soil and surface disturbance ecological characteristics; constructing a regional ecological feature vector based on the multi-dimensional ecological environment characteristics; inputting the regional ecological feature vector into a pre-constructed and trained ecological environment quality assessment model group to obtain ecological environment quality assessment results for multiple ecological zones; and spatially stitching the ecological environment quality assessment results of each ecological zone to generate the target ecological environment quality assessment result. This invention, through techniques such as consistent preprocessing of multi-source remote sensing data, extraction of multi-dimensional ecological environment characteristics, and ecological quality assessment based on regional adaptive ecological environment quality weight modeling, not only integrates multi-source remote sensing data but also takes into account regional differences, enabling continuous, refined, and comprehensive assessment of ecological environment quality, thereby providing scientific support for ecological environment protection and resource management.

[0077] Figure 2 The structural block diagram of an ecological environment quality assessment device that combines multi-source remote sensing data is shown in the embodiment of this application.

[0078] The ecological environment quality assessment device combining multi-source remote sensing data provided in this application includes the following functional modules: The acquisition module 201 is used to acquire multi-source remote sensing data containing ecological and environmental characteristics, wherein the multi-source remote sensing data includes: optical remote sensing data, radar remote sensing data and terrain data; Preprocessing module 202 is used to preprocess the multi-source remote sensing data; Extraction module 203 is used to extract multi-dimensional ecological and environmental features from preprocessed multi-source remote sensing data, wherein the multi-dimensional ecological and environmental features include: vegetation ecological features, water body and humidity ecological features, thermal environment ecological features, and soil and surface disturbance ecological features. Construction module 204 is used to construct a regional ecological feature vector based on the multidimensional ecological environment characteristics; The evaluation module 205 is used to input the regional ecological feature vector into a pre-constructed and trained ecological environment quality evaluation model group to obtain ecological environment quality evaluation results for multiple ecological zones; wherein, the ecological environment quality evaluation model group includes multiple ecological zone sub-models constructed with adaptive ecological environment quality weights for each ecological zone; The generation module 206 is used to spatially stitch together the ecological environment quality assessment results of each ecological zone to generate the target ecological environment quality assessment results.

[0079] Optionally, the preprocessing module includes: The first submodule is used to perform radiometric calibration and atmospheric correction on optical remote sensing data containing ecological and environmental characteristics, so as to convert optical remote sensing images into surface reflectance data. The second submodule is used to perform geometric correction and projection unification processing on the optical remote sensing data so that the optical remote sensing images are consistent in spatial position. The third submodule is used to perform terrain correction on the optical remote sensing data based on the terrain data, and to perform spatial resolution unification processing on the terrain-corrected optical remote sensing data. The fourth submodule is used to detect and repair optical pixels contaminated by clouds or shadows in radar remote sensing data containing ecological and environmental characteristics.

[0080] Optionally, the extraction module includes: The fifth submodule is used to extract vegetation coverage and vegetation growth status from preprocessed multi-source remote sensing data as vegetation ecological characteristics. The sixth submodule is used to extract water body identification information and surface humidity from preprocessed multi-source remote sensing data as ecological characteristics of water bodies and humidity, depending on different underlying surface conditions. The seventh submodule is used to extract land surface temperature as a thermal environment ecological feature from preprocessed multi-source remote sensing data. The eighth submodule is used to extract surface aridity and bare land index from preprocessed multi-source remote sensing data as ecological characteristics of soil and surface disturbance.

[0081] Optionally, the building module is specifically used for: The multidimensional ecological environment characteristics are standardized. A regional ecological feature vector is constructed based on the standardized multidimensional ecological environment characteristics.

[0082] Optionally, the ecological environment quality assessment model set is trained and generated in the following manner: Based on land use data and ecological function zoning results, the study area was divided into multiple ecological zones; In each ecological zone, a sample set is constructed with multidimensional ecological environment characteristics as input and historical ecological environment quality reference values ​​as output. In each ecological zone, a predetermined number of principal component features are extracted from the multidimensional ecological environment features based on principal component analysis; Establish a regression model based on principal component characteristics within each ecological zone; For each regression model, the regression coefficients of the regression model are mapped to normalized weights, and the normalized weights are subject to soft constraint correction. Within each ecological zone, a regional adaptive ecological environment quality index is constructed based on the normalized weights after soft constraint correction.

[0083] The embodiments provided in this application Figure 2 The ecological environment quality assessment device shown, which combines multi-source remote sensing data, can achieve... Figure 1 The various processes and beneficial effects implemented in the method implementation examples will not be repeated here to avoid repetition.

[0084] This invention also provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus.

[0085] Memory, used to store computer programs; When the processor executes the program stored in the memory, it implements the ecological environment quality assessment method combining multi-source remote sensing data as shown in the above method embodiments.

[0086] The communication bus mentioned above can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc.

[0087] The communication interface is used for communication between the aforementioned terminal and other devices.

[0088] The memory may include random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Optionally, the memory may also be at least one storage device located remotely from the aforementioned processor.

[0089] In another embodiment of the present invention, a computer-readable storage medium is also provided, which stores instructions that, when executed on an electronic device, cause the electronic device to implement the ecological environment quality assessment method combining multi-source remote sensing data as described in any of the above embodiments.

[0090] In another embodiment of the present invention, a computer program product containing instructions is also provided, which, when run on an electronic device, causes the electronic device to implement any of the above embodiments of the method for assessing the ecological environment quality by combining multi-source remote sensing data.

[0091] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0092] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for assessing ecological environment quality by combining multi-source remote sensing data, characterized in that, The method includes: Acquire multi-source remote sensing data containing ecological and environmental characteristics, wherein the multi-source remote sensing data includes: optical remote sensing data, radar remote sensing data, and topographic data; The multi-source remote sensing data is preprocessed; Multidimensional ecological and environmental features are extracted from preprocessed multi-source remote sensing data. These features include vegetation ecological features, water body and humidity ecological features, thermal environment ecological features, and soil and surface disturbance ecological features. Construct a regional ecological feature vector based on the aforementioned multidimensional ecological environment characteristics; The regional ecological feature vector is input into a pre-constructed and trained ecological environment quality assessment model set to obtain ecological environment quality assessment results for multiple ecological zones; wherein, the ecological environment quality assessment model set includes multiple ecological zone sub-models constructed with adaptive ecological environment quality weights for each ecological zone; The ecological environment quality assessment results of each ecological zone are spatially stitched together to generate the target ecological environment quality assessment result.

2. The method according to claim 1, characterized in that, The steps for preprocessing the multi-source remote sensing data include: Radiometric calibration and atmospheric correction are performed on optical remote sensing data containing ecological and environmental characteristics to convert optical remote sensing images into surface reflectance data; The optical remote sensing data is subjected to geometric correction and projection unification processing to ensure that the optical remote sensing images are consistent in spatial location. Based on the terrain data, the optical remote sensing data is subjected to terrain correction, and the spatial resolution of the terrain-corrected optical remote sensing data is uniformly processed. Detect and repair optical pixels contaminated by clouds or shadows in radar remote sensing data containing ecological and environmental characteristics.

3. The method according to claim 1, characterized in that, The steps for extracting multidimensional ecological and environmental features from preprocessed multi-source remote sensing data include: Vegetation coverage and vegetation growth status are extracted from preprocessed multi-source remote sensing data as vegetation ecological characteristics. For different underlying surface conditions, water body identification information and surface humidity are extracted from preprocessed multi-source remote sensing data as ecological characteristics of water bodies and humidity. Surface temperature is extracted from preprocessed multi-source remote sensing data as a thermal environment ecological feature. Surface aridity and bare land index were extracted from preprocessed multi-source remote sensing data as ecological characteristics of soil and surface disturbance.

4. The method according to claim 1, characterized in that, The steps for constructing a regional ecological feature vector based on the aforementioned multidimensional ecological environment characteristics include: The multidimensional ecological environment characteristics are standardized. A regional ecological feature vector is constructed based on the standardized multidimensional ecological environment characteristics.

5. The method according to claim 1, characterized in that, The ecological environment quality assessment model set was trained and generated in the following manner: Based on land use data and ecological function zoning results, the study area was divided into multiple ecological zones; In each ecological zone, a sample set is constructed with multidimensional ecological environment characteristics as input and historical ecological environment quality reference values ​​as output. In each ecological zone, a predetermined number of principal component features are extracted from the multidimensional ecological environment features based on principal component analysis; Establish a regression model based on principal component characteristics within each ecological zone; For each regression model, the regression coefficients of the regression model are mapped to normalized weights, and the normalized weights are subject to soft constraint correction. Within each ecological zone, a regional adaptive ecological environment quality index is constructed based on the normalized weights after soft constraint correction.

6. An ecological environment quality assessment device combining multi-source remote sensing data, characterized in that, The device includes: The acquisition module is used to acquire multi-source remote sensing data containing ecological and environmental characteristics, wherein the multi-source remote sensing data includes: optical remote sensing data, radar remote sensing data, and terrain data; The preprocessing module is used to preprocess the multi-source remote sensing data; The extraction module is used to extract multidimensional ecological and environmental features from preprocessed multi-source remote sensing data. The multidimensional ecological and environmental features include: vegetation ecological features, water body and humidity ecological features, thermal environment ecological features, and soil and surface disturbance ecological features. The construction module is used to construct a regional ecological feature vector based on the multidimensional ecological environment characteristics; The evaluation module is used to input the regional ecological feature vector into a pre-constructed and trained ecological environment quality evaluation model group to obtain ecological environment quality evaluation results for multiple ecological zones; wherein, the ecological environment quality evaluation model group includes multiple ecological zone sub-models constructed with adaptive ecological environment quality weights for each ecological zone; The generation module is used to spatially stitch together the ecological environment quality assessment results of each ecological zone to generate the target ecological environment quality assessment result.

7. The apparatus according to claim 6, characterized in that, The preprocessing module includes: The first submodule is used to perform radiometric calibration and atmospheric correction on optical remote sensing data containing ecological and environmental characteristics, so as to convert optical remote sensing images into surface reflectance data. The second submodule is used to perform geometric correction and projection unification processing on the optical remote sensing data to make the optical remote sensing images consistent in spatial position. The third submodule is used to perform terrain correction on the optical remote sensing data based on the terrain data, and to perform spatial resolution unification processing on the terrain-corrected optical remote sensing data. The fourth submodule is used to detect and repair optical pixels contaminated by clouds or shadows in radar remote sensing data containing ecological and environmental characteristics.

8. The apparatus according to claim 6, characterized in that, The extraction module includes: The fifth submodule is used to extract vegetation coverage and vegetation growth status from preprocessed multi-source remote sensing data as vegetation ecological characteristics. The sixth submodule is used to extract water body identification information and surface humidity from preprocessed multi-source remote sensing data as ecological characteristics of water bodies and humidity, depending on different underlying surface conditions. The seventh submodule is used to extract land surface temperature as a thermal environment ecological feature from preprocessed multi-source remote sensing data. The eighth submodule is used to extract surface aridity and bare land index from preprocessed multi-source remote sensing data as ecological characteristics of soil and surface disturbance.

9. The apparatus according to claim 6, characterized in that, The building module is specifically used for: The multidimensional ecological environment characteristics are standardized. A regional ecological feature vector is constructed based on the standardized multidimensional ecological environment characteristics.

10. An electronic device, characterized in that, It includes a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus; Memory, used to store computer programs; The processor, when executing a program stored in memory, implements the ecological environment quality assessment method combining multi-source remote sensing data as described in any one of claims 1-5.

Images